A5086174310- Dementia and Cognitive Impairment Research
- Machine Learning in Healthcare
- Functional Brain Connectivity Studies
- Neurobiology of Language and Bilingualism
- Alzheimer's disease research and treatments
- Gut microbiota and health
- Dermatology and Skin Diseases
- Advanced Clustering Algorithms Research
- Health, Environment, Cognitive Aging
- Health Systems, Economic Evaluations, Quality of Life
- Advanced Neuroimaging Techniques and Applications
- Chronic Disease Management Strategies
- Genetic Associations and Epidemiology
- Genomics and Rare Diseases
- Neurological disorders and treatments
- EEG and Brain-Computer Interfaces
- Neurological Disease Mechanisms and Treatments
- Parkinson's Disease Mechanisms and Treatments
- Medical Image Segmentation Techniques
- Multiple Sclerosis Research Studies
- Metaheuristic Optimization Algorithms Research
- Evolutionary Algorithms and Applications
- Systemic Lupus Erythematosus Research
- Bioinformatics and Genomic Networks
- Electron and X-Ray Spectroscopy Techniques
University College London
2021-2025
Sci-Tech Daresbury
2020-2021
University of Manchester
2018-2020
Manchester Metropolitan University
2020
Abstract Alterations in the human microbiome have been observed a variety of conditions such as asthma, gingivitis, dermatitis and cancer, much remains to be learned about links between health. The fusion artificial intelligence with rich datasets can offer an improved understanding microbiome’s role To gain actionable insights it is essential consider both predictive power transparency models by providing explanations for predictions. We combine collection leg skin samples from two healthy...
Progressive disorders are highly heterogeneous. Symptom-based clinical classification of these may not reflect the underlying pathobiology. Data-driven subtyping and staging patients has potential to disentangle complex spatiotemporal patterns disease progression. Tools that enable this in high demand from treatment-development communities. Here we describe pySuStaIn software package, a Python-based implementation Subtype Stage Inference (SuStaIn) algorithm. SuStaIn unravels complexity...
Abstract To better understand the pathological and phenotypic heterogeneity of progressive supranuclear palsy links between two, we applied a novel unsupervised machine learning algorithm (Subtype Stage Inference) to largest MRI data set date people with clinically diagnosed (including palsy–Richardson variant syndromes). Our cohort is comprised 426 cases, which 367 had at least one follow-up scan, 290 controls. Of 357 were palsy–Richardson, 52 palsy–cortical (progressive palsy–frontal,...
Abstract Cortical atrophy and aggregates of misfolded tau proteins are key hallmarks Alzheimer’s disease. Computational models that simulate the propagation pathogens between connected brain regions have been used to elucidate mechanistic information about spread these disease biomarkers, such as epicentres spreading rates. However, connectomes substrates for known contain modality-specific false positive negative connections, influenced by biases inherent different methods estimating...
Although the corticobasal syndrome was originally most closely linked with pathology of degeneration, 2013 Armstrong clinical diagnostic criteria, without addition aetiology-specific biomarkers, have limited positive predictive value for identifying degeneration in life. Autopsy studies demonstrate considerable pathological heterogeneity syndrome, accounting only ∼50% clinically diagnosed individuals. Individualized disease stage and progression modelling brain changes may utility predicting...
Parkinson's disease is the second most common neurodegenerative disease. Despite this, there are no robust biomarkers to predict progression, and understanding of mechanisms limited. We used Subtype Stage Inference algorithm characterize heterogeneity in terms spatiotemporal subtypes macroscopic atrophy detectable on T1-weighted MRI-a successful approach other diseases. trained model covariate-adjusted cortical thicknesses subcortical volumes from largest known MRI dataset disease, Enhancing...
Correlative light and volume electron microscopy (vCLEM) is a powerful imaging technique that enables the visualisation of fluorescently labelled proteins within their ultrastructural context on subcellular level. Currently, expert microscopists align vCLEM acquisitions using time-consuming subjective manual methods. This paper presents CLEM-Reg, an algorithm automates 3D alignment datasets by leveraging probabilistic point cloud registration techniques. These clouds are derived from...
Heterogeneity in Alzheimer's disease progression contributes to the ongoing failure demonstrate efficacy of putative disease-modifying therapeutics that have been trialed over past two decades. Any treatment effect present a subgroup trial participants (responders) can be diluted by non-responders who ideally should screened out trial. How identify (screen-in) most likely potential responders is an important question still without answer. Here, we pilot computational screening tool leverages...
Synthetic datasets play an important role in evaluating clustering algorithms, as they can help shed light on consistent biases, strengths, and weaknesses of particular techniques, thereby supporting sound conclusions. Despite this, there is a surprisingly small set established benchmark data, many these are currently handcrafted. Even then, their difficulty typically not quantified or considered, limiting the ability to interpret algorithmic performance datasets. Here, we introduce HAWKS,...
The primary progressive aphasias are rare, language-led dementias, with three main variants: semantic, non-fluent/agrammatic, and logopenic. Whilst semantic variant has a clear neuroanatomical profile, the non-fluent/agrammatic logopenic variants difficult to discriminate from neuroimaging. Previous phenotype-driven studies have characterised profiles of each on MRI. In this work we used machine learning algorithm known as SuStaIn discover data-driven "subtype" progression performed an...
Undetected biological heterogeneity adversely impacts trials in Alzheimer's disease because rate of cognitive decline - and perhaps response to treatment differs subgroups. Recent results show that data-driven approaches can unravel the progression. The resulting stratification is yet be leveraged clinical trials. Investigate whether image-based progression modelling could identify baseline a trial, these subgroups have prognostic or predictive value. Screening data from Anti-Amyloid...
Comprehensive benchmarking of clustering algorithms is rendered difficult by two key factors: (i)~the elusiveness a unique mathematical definition this unsupervised learning approach and (ii)~dependencies between the generating models or criteria adopted some indices for internal cluster validation. Consequently, there no consensus regarding best practice rigorous benchmarking, whether possible at all outside context given application. Here, we argue that synthetic datasets must continue to...
Abstract Alterations in the human microbiome have been observed a variety of conditions such has asthma, gingivitis, dermatitis and cancer, much remains to be learned about links between health. The fusion artificial intelligence with rich datasets can offer an improved understanding microbiome’s role our To gain actionable insights it is essential consider both predictive power transparency models by providing explanations for predictions. We combine effort collecting corpus leg skin...
To identify imaging subtypes of the cortico-basal syndrome (CBS) based solely on a data-driven assessment MRI atrophy patterns, and investigate whether these provide information underlying pathology.
Abstract The primary progressive aphasias are rare, language-led dementias, with three main variants: semantic, non-fluent/agrammatic, and logopenic. Whilst semantic variant has a clear neuroanatomical profile, the non-fluent/agrammatic logopenic variants difficult to discriminate from neuroimaging. Previous phenotype-driven studies have characterised profiles of each on MRI. In this work we used machine learning algorithm known as SuStaIn discover data-driven “subtype” progression performed...
Alzheimer's disease (AD) exhibits substantial clinical and biological heterogeneity, complicating efforts in treatment intervention development. While new computational methods offer insights into AD progression, the reproducibility of these subtypes across datasets remains understudied, particularly concerning robustness subtype definitions when validated on diverse databases. This study evaluates consistency progression identified by Subtype Stage Inference (SuStaIn) algorithm using...
ABSTRACT Heterogeneity in Alzheimer’s disease progression contributes to the ongoing failure demonstrate efficacy of putative disease-modifying therapeutics that have been trialled over past two decades. Any treatment effect present a subgroup trial participants (responders) can be diluted by non-responders who ideally should screened out trial. How identify (screen-in) most likely potential responders is an important question still without answer. Here we pilot computational screening tool...
Abstract Progressive disorders are highly heterogeneous. Symptom-based clinical classification of these may not reflect the underlying pathobiology. Data-driven subtyping and staging patients has potential to disentangle complex spatiotemporal patterns disease progression. Tools that enable this in high demand from treatment-development communities. Here we describe pySuStaIn software package, a Python-based implementation Subtype Stage Inference (SuStaIn) algorithm. SuStaIn unravels...
Introduction: Machine learning (ML) has been extremely successful in identifying key features from high-dimensional datasets and executing complicated tasks with human expert levels of accuracy or greater. Methods: We summarize critically evaluate current applications ML dementia research highlight directions for future research. Results: present an overview algorithms most frequently used opportunities the use clinical practice, experimental medicine, trials. discuss issues reproducibility,...
Machine learning models offer the potential to understand diverse datasets in a data-driven way, powering insights into individual disease experiences and ensuring equitable healthcare. In this study, we explore Bayesian inference for characterising symptom sequences, associated modelling challenges. We adapted Mallows model account partial rankings right-censored data, employing custom MCMC fitting. Our evaluation, encompassing synthetic data primary progressive aphasia dataset, highlights...
Abstract Background The preclinical phase of Alzheimer’s disease (AD), where pathology slowly accumulates years before cognitive impairment becomes apparent, could offer a treatment window with the greatest potential to preserve function downstream pathological processes gather momentum. Characterizing when biomarker trajectories deviate from normal ageing, and heterogeneity therein, facilitate targeted trial recruitment improved biomarker‐based evidence modification. However, reliably...
Abstract Background Postmortem diagnosis of Alzheimer’s and related pathologies remains the gold standard. Advanced neuroimaging techniques enable assessment some in living subjects, but antemortem signatures are often defined post hoc postmortem‐defined groups. Here we flip this estimate data‐driven pathology using Subtype Stage Inference (SuStaIn), then predict postmortem classification. SuStaIn is a machine learning algorithm that jointly estimates subtype clusters disease progression to...
Abstract Background Primary progressive aphasia (PPA) is an atypical neurodegenerative dementia with three main clinical variants: semantic (svPPA), non‐fluent/agrammatic (nfvPPA) and logopenic (lvPPA). While svPPA typically associated left anterior temporal lobe atrophy, neuroimaging findings in nfvPPA, lvPPA PPA not otherwise specified (PPA‐nos) are more variable. We therefore applied unsupervised machine learning to one of the largest databases magnetic resonance imaging (MRI) scans ever...